d3rlpy.algos.ConstantEpsilonGreedy¶
- class d3rlpy.algos.ConstantEpsilonGreedy(epsilon)[source]¶
\(\epsilon\)-greedy explorer with constant \(\epsilon\).
- Parameters
epsilon (float) – the constant \(\epsilon\).
Methods
- sample(algo, x, step)[source]¶
- Parameters
algo (d3rlpy.interface.QLearningAlgoProtocol) –
x (Union[numpy.ndarray[Any, numpy.dtype[Any]], Sequence[numpy.ndarray[Any, numpy.dtype[Any]]]]) –
step (int) –
- Return type
numpy.ndarray[Any, numpy.dtype[Any]]